---
title: Twisted Web Immanants in Grassmannians
url: https://www.emergentmind.com/topics/twisted-web-immanants
type: topic
---

# Twisted Web Immanants in Grassmannians

Searching arXiv for the primary and related papers on twisted web immanants and immanant generalizations.
{"query":"arXiv 2507.15211 twisted web immanants Grassmannian cluster algebras", "max_results": 5}
Searching arXiv for related work on immanants, web immanants, and diagram-algebra generalizations.
Twisted web immanants are the twist-transformed analogues of ordinary web immanants in the Grassmannian web formalism. In the formulation of “Twists, Higher Dimer Covers, and Web Duality for Grassmannian Cluster Algebras,” they are not introduced as a separate abstract object via a new immanant isomorphism; instead, they arise from a twisted higher boundary measurement map that replaces the ordinary edge-weighted \(r\)-dimer sum by a face-weighted \(r\)-dimer sum, so that pairing against this web-valued object computes the twist \(\tau(f)\) of a Grassmannian function \(f\) [2507.15211]. The resulting theory sits at the intersection of \(\mathbb C[\Gr(k,n)]\), \(\mathrm{SL}_r\) web invariants, reduced top-cell plabic graphs, \(r\)-dimer covers, and Laurent expansions in Plücker coordinates.

## 1. Ambient Grassmannian and web-immanant framework

The ambient ring is the Grassmannian coordinate ring
\[
\mathbb C[\Gr(k,n)],
\]
where \(\Gr(k,n)\) is the affine cone over the Plücker embedding of the Grassmannian of \(k\)-planes in \(\mathbb C^n\). It is generated by Plücker coordinates \(\Delta_I\), \(I\in \binom{[n]}{k}\), and carries its natural \(\mathbb N^n\)-grading, with \(\Delta_I\) of degree \(\delta^I\). A reduced top-cell plabic graph \(G\) of type \((k,n)\) determines an initial seed for the cluster algebra structure: each face \(\mathbf f\) is labeled by a \(k\)-subset \(I_{\mathbf f}\), and the corresponding cluster variable is \(\Delta_{I_{\mathbf f}}\) [2507.15211].

For \(\lambda\in \mathbb N^n\) with \(|\lambda|=kr\), the relevant invariant space is
\[
\mathcal W_\lambda(\mathbb C^r)= \operatorname{Hom}_{\mathrm{SL}_r}\!\left(\bigotimes_{i=1}^n \bigwedge^{\lambda_i}\mathbb C^r,\mathbb C\right),
\]
spanned by invariants attached to planar \(\mathrm{SL}_r\) webs. Fraser–Lam–Le’s immanant formalism identifies the graded piece \(\mathbb C[\Gr(k,n)]_\lambda\) with the dual web space through the isomorphism
\[
\Imm:(\mathcal W_\lambda(\mathbb C^r))^* \xrightarrow{\sim} \mathbb C[\Gr(k,n)]_\lambda,
\]
equivalently through a perfect pairing
\[
\langle \cdot,\cdot\rangle: \mathcal W_\lambda(\mathbb C^r)\otimes \mathbb C[\Gr(k,n)]_\lambda\to \mathbb C.
\]
In this setting, an ordinary web immanant is the Grassmannian function corresponding to a functional on the web space.

The combinatorial engine is the set \(\mathcal D_{r,\lambda}(G)\) of \(r\)-dimer covers of a reduced top-cell plabic graph \(G\). An \(r\)-dimer cover \(D\) is a multiset of edges such that every internal vertex is incident to exactly \(r\) edges and boundary vertex \(i\) is used with multiplicity \(\lambda_i\). Each such \(D\) determines an \(\mathrm{SL}_r\) web, written \(\mathbf D\in \mathcal W_\lambda(\mathbb C^r)\). For a network \(N\) on \(G\) with edge weights \(\operatorname{wt}(\mathbf e)\), the ordinary higher boundary measurement is
\[
\Web_r(N;\lambda) = \sum_{D\in \mathcal D_{r,\lambda}(G)} \ewt_N(D)\,\mathbf D,
\qquad
\ewt_N(D)=\prod_{\mathbf e\in D}\operatorname{wt}(\mathbf e),
\]
and it satisfies
\[
\langle \Web_r(N;\lambda),f\rangle = f(\widehat X(N)).
\]
Twisted web immanants are defined relative to this ordinary web-immanant story: the coefficient system \(\langle \mathbf D,f\rangle\) is unchanged, but the monomials are altered by replacing edge weights with face-weight Laurent monomials.

## 2. Twisted higher boundary measurement and the defining formula

The fundamental new ingredient is the face-weight of an \(r\)-dimer cover. If \(\face\) is a face of \(G\), let \(I_\face\) be its \(k\)-subset label, let \(W_\face\) denote the number of white vertices bordering \(\face\), and let \(D_\face\) denote the number of non-boundary edges of \(\face\) used in \(D\). Then the face weight is
\[
\fwt_G(D)\coloneqq\prod_{\face\in F(G)}\Delta_{I_\face}^{\,rW_\face-D_\face-r}.
\]
This is a Laurent monomial in the Plücker coordinates of the seed attached to \(G\) [2507.15211].

The central theorem is the twist formula
\[
\tau(f)=\sum_{D\in \mathcal D_{r,\lambda}(G)} \langle \mathbf D,f\rangle\, \fwt_G(D),
\]
valid for \(f\in \mathbb C[\Gr(k,n)]_\lambda\) with \(|\lambda|=kr\). This formula is the paper’s operational definition of twisted web immanants: the coefficients are the usual Fraser–Lam–Le pairing coefficients, while the basis monomials are the face-weight Laurent monomials rather than the ordinary edge-weight monomials [2507.15211].

Evaluating the Laurent monomial at a network point \(\widehat X(N)\) gives
\[
\fwt_N(D)\coloneqq \fwt_G(D)(\widehat X(N)),
\]
and the twisted higher boundary measurement map is
\[
\Web_r^\tau(N;\lambda)\coloneqq \sum_{D\in \mathcal D_{r,\lambda}(G)} \fwt_N(D)\mathbf D.
\]
Its defining property is
\[
\langle \Web_r^\tau(N;\lambda),f\rangle = (\tau(f))(\widehat X(N)).
\]
Thus \(\Web_r^\tau(N;\lambda)\) plays for \(\tau(f)\) exactly the role that \(\Web_r(N;\lambda)\) plays for \(f\). In particular, the twisted web-immanant formalism is not a replacement of the Fraser–Lam–Le pairing, but a twist-compatible reweighting of the universal \(r\)-dimer generating series.

The case \(r=1\) recovers the twisted Plücker formula
\[
\tau(\Delta_I)= \sum_{D\in\mathcal D_{1,\delta^I}(G)}\fwt_G(D),
\]
so the higher-\(r\) theory is a direct extension of the Marsh–Scott and Elkin–Musiker–Wright dimer formulas from Plücker coordinates to arbitrary homogeneous Grassmannian functions.

## 3. Web duality and dual-basis extraction

A distinctive simplification occurs when the dual basis under the Fraser–Lam–Le pairing is again realized by web invariants. When \(n=kr\) and \(\lambda=(1,\dots,1)\), one has
\[
\mathbb C[\Gr(k,n)]_\lambda \cong \mathcal W_\lambda(\mathbb C^k),
\]
so the pairing yields a duality between \(\mathcal W_\lambda(\mathbb C^r)\) and \(\mathcal W_\lambda(\mathbb C^k)\). In this setting, “web duality” means that the basis dual under the pairing to a web basis on one side is again given by web invariants on the other side, often indexed by transpose tableaux [2507.15211].

The paper proves this duality for a large family of \(\mathrm{SL}_3\) and \(\mathrm{SL}_4\) webs at \(n=12\). More precisely, for the bases \(\mathcal B\) and \(\mathcal B^*\) appearing in the \((3,12)\) and \((4,12)\) situations, the pairing satisfies
\[
\langle \mathbf W_i,\mathbf X_j\rangle
=
\begin{cases}
\operatorname{sign}(X_j), & i=j,\\
0, & i\ne j,
\end{cases}
\]
and the corresponding tableaux satisfy \(T(W_i)=T(X_i)^t\). This gives a concrete dual-web description of twisted web immanants in those cases.

The resulting twisted formula is particularly transparent. If \(\mathbf Y\in \mathbb C[\Gr(k,12)]_\lambda\) is the invariant of an \(\mathrm{SL}_k\)-basis web \(Y\), and \(Y^*\) denotes the dual basis web under the pairing, then
\[
\tau(\mathbf Y)= \operatorname{sign}(Y)\sum_{D\in \mathcal D_r(G)} C_{Y^*}^D\,\fwt_G(D),
\]
where \(C_{Y^*}^D\) is the coefficient of \(\mathbf Y^*\) when \(\mathbf D\) is expanded in the chosen web basis. In these cases, computing the twisted web immanant reduces to extracting a single dual-web coefficient from each \(r\)-weblike subgraph.

The theory is not uniform without qualifications. The paper identifies one exceptional \(\mathrm{SL}_4\) dual basis element in the \((3,12)\)/\((4,12)\) story: the dual to \(X_{16}\) is not a single basis web in the chosen \(\mathrm{SL}_4\) basis but a difference of two webs, the “benzene case.” Even there, however, the twisted expansion formula still applies.

## 4. Cluster twists, Laurent expansions, and explicit computations

The principal application is to Grassmannian cluster algebra twists. By combining the twisted higher boundary measurement map with web duality, the paper recovers and extends formulas of Elkin–Musiker–Wright for twists of certain cluster variables. The mechanism is uniform: if \(f\) is a homogeneous Grassmannian function represented by a web invariant and its dual web is known, then \(\tau(f)\) is obtained by summing the face weights of precisely those \(r\)-dimer covers whose web expansions contain the dual web with nonzero coefficient [2507.15211].

A model example is the \(\mathrm{SL}_3\) basis web invariant \(\mathbf X_{28}^{\mathrm{FP}}\in \mathbb C[\Gr(3,12)]_{(1^{12})}\). The paper gives an explicit Plücker expansion for \(\mathbf X_{28}^{\mathrm{FP}}\), an explicit formula for its twist, and then a Laurent expansion in the initial seed. It further explains that this Laurent polynomial matches the face weights of the only two quadruple dimer covers of the chosen reduced top-cell plabic graph whose associated \(4\)-weblike subgraphs expand, under \(\mathrm{SL}_4\) skein relations, to include the dual web \(W_{28}\); in each case, \(W_{28}\) appears with coefficient \(1\). This example makes the coefficient-extraction principle completely explicit.

The cluster-theoretic significance is broader than isolated examples. The paper presents evidence supporting conjectures of Fomin–Pylyavskyy and one by Cheung–Dechant–He–Heyes–Hirst–Li, and proves an enumeration theorem for arborizable indecomposable \(\mathrm{SL}_3\) webs of Plücker degree \(4\),
\[
T_{3,n,4}=288\binom n9 + 400\binom n{10} + 264\binom n{11} + 52\binom n{12}.
\]
This is evidence that the web-theoretic picture underlying twisted web immanants aligns with conjectural classifications of low-degree cluster variables in \(\mathbb C[\Gr(3,n)]\).

At the same time, the scope remains specific. The paper proves general twist formulas for arbitrary homogeneous \(f\in \mathbb C[\Gr(k,n)]_\lambda\), but the most explicit dual-web coefficient formulas are obtained in the settings where web duality has been established, notably for many \(\mathrm{SL}_3\) and \(\mathrm{SL}_4\) webs at \(n=12\).

## 5. Relation to classical and diagram-algebra immanants

Twisted web immanants belong to a larger ecosystem of immanant-like constructions, but the relationship to more classical theories is indirect and comparative rather than definitional. The sharpest contrast is with ordinary character immanants of the symmetric group,
\[
d_\chi(A)=\sum_{\sigma\in S_n}\chi(\sigma)\prod_{i=1}^n a_{i,\sigma(i)}.
\]
For these classical immanants, vanishing on skew-symmetric complex matrices is completely classified: for even \(n\), vanishing occurs exactly for those \(\chi^\lambda\) indexed by indestructible diagrams under recursive domino rim-hook removal, and this is equivalent to vanishing on all permutations with only even cycles, already equivalently on the fixed-point-free involution class \((2^{n/2})\) [2402.05710]. This provides a precise benchmark for the “vanishing on alternating matrices” problem, but it does not itself address twisted web immanants.

A second comparison comes from the recombinant, an immanant-like function defined using partition algebra characters and a sum over all partition diagrams rather than over permutations:
\[
\operatorname{Rec}^{\lambda}(A)=\sum_{d\in P_n(r)}\chi^\lambda_{P_n(r)}(d)\,{}_d a_{i,j}.
\]
Its relevance lies in the structural analogy: it shows how an immanant-type function can be built from a larger diagram algebra, and how a quotient or top-propagation sector recovers the classical immanant [2309.16647]. Twisted web immanants share the passage from permutation combinatorics to a richer diagrammatic setting, but their coefficients come from web expansions and the Fraser–Lam–Le pairing, not from partition algebra characters.

Temperley–Lieb immanants furnish a closer diagrammatic analogue on the rank-two side. They are defined from the Temperley–Lieb basis and admit complementary-minor expansions of the form
\[
\Delta_{I,J}\Delta_{\overline I,\overline J}
=
\sum_{w\ \mathrm{compatible\ with}\ (I,J)} \imm_w,
\]
with compatibility encoded by planar non-crossing matchings and colorings [2303.17004]. This is not a theorem about twisted web immanants, but it is one of the clearest lower-rank prototypes for how a planar diagram basis can control an immanant family through minor expansions.

A further comparison comes from Kazhdan–Lusztig immanants and dual canonical basis elements of \(\mathbb C[SL_m]\). For \(1324\)- and \(2143\)-avoiding permutations, certain Kazhdan–Lusztig immanants reduce to signed determinants of support-restricted matrices, and the corresponding dual canonical basis elements are \(k\)-positive under an explicit square-size condition [2106.09150]. This suggests a broader pattern: complicated immanants may become tractable on distinguished combinatorial classes. Twisted web immanants realize this pattern in a different way, through face-weighted \(r\)-dimer sums and dual-web extraction rather than support-restricted determinants.

## 6. Scope, misconceptions, and interpretive boundaries

Twisted web immanants are not ordinary \(S_n\)-character immanants. Their coefficients are not given by irreducible characters of the symmetric group, and the relevant indexing data are webs, \(r\)-dimer covers, and dual web bases rather than partitions alone. For that reason, classical results on ordinary immanants provide comparison principles but not direct formulas.

They are also not introduced by a separate new immanant isomorphism. The defining mechanism is the twisted higher boundary measurement map together with the existing Fraser–Lam–Le pairing. The same coefficient \(\langle \mathbf D,f\rangle\) that controls ordinary web immanants controls the twisted version; what changes is the replacement of edge-weight monomials by face-weight Laurent monomials in the initial Plücker seed [2507.15211].

A common overstatement is to identify web duality with a completely general theorem. The paper proves that web duality continues to hold for a large set of \(\mathrm{SL}_3\) and \(\mathrm{SL}_4\) webs, and it verifies tableau transposition in those cases, but it does not establish the phenomenon for all \(n=kr\). The benzene exception likewise shows that even within the proved range, the dual basis need not always be realized by a single basis web.

The comparison with other immanant generalizations must also be qualified. The vanishing-immanant classification on alternating matrices [2402.05710], the recombinant construction from partition algebra characters [2309.16647], the Temperley–Lieb theory [2303.17004], and the \(k\)-positivity results for Kazhdan–Lusztig immanants [2106.09150] do not mention twisted web immanants directly. Their relevance is that they provide explicit models for three recurrent themes: diagram-algebra enlargement, basis-dependent coefficient systems, and reduction of immanants to tractable combinatorial or determinantal formulas.

A plausible implication is that future work on twisted web immanants may continue to proceed by identifying the correct dual basis, the correct diagrammatic coefficient-extraction rule, and the correct test families of dimer covers or support patterns. What is already established is narrower and more precise: twisted web immanants, in the current literature, are Grassmannian twist formulas expressed through face-weighted \(r\)-dimer sums and the Fraser–Lam–Le pairing, with especially explicit forms when web duality is known.

Source: https://www.emergentmind.com/topics/twisted-web-immanants